Video target tracking method based on SVM and Mean-Shift

Abstract

The invention discloses a video target tracking method based on SVM and Mean-Shift. The method comprise the following steps: selecting target scope and background scope in a first frame image; training an SVM classifier based on pixel data in the first frame image; obtaining a confidence map comprising target and background two parts by utilizing the SVM classifier pixel and data in the background scope at the same position in the next frame image; obtaining a target center position within the confidence map scope by utilizing the Mean-Shift mean shift method; zooming the size of a target frame with the 10% proportion at the target position and comparing the zooming result with a target in the previous frame, the most similar one being the size of the final target; and training new SVM classifier based on the data of the next frame, repeating the step 3 to circle continuously until tracking to the last frame, and the tracking process is finished. The video target tracking method is based on the SVM training classifier and the Mean-Shift method, so that the video target tracking method is good in real-time performance, accuracy and robustness, and is appropriate for the dynamic background and the tracking of nonrigid targets.

Description

technical field [0001] The invention relates to a video tracking method, in particular to a video target tracking method based on SVM and Mean-Shift. Background technique [0002] With the development of computer vision technology, moving video target tracking technology has become an important topic in the fields of weapon guidance, pattern recognition, and computer vision. Because moving video target tracking technology has broad application prospects in both military and civilian fields, many scholars at home and abroad are engaged in the research of this subject and have proposed many classic target tracking methods. Moving video target tracking organically combines image processing, computer vision and information science to form a technology that can automatically identify targets from video images in real time, extract target location information, and automatically track targets. [0003] Video tracking is the process of identifying the target of interest and analyzi...

AI Technical Summary

Problems solved by technology

However, when the target has obvious scale changes, especially when the target size gradually increases beyond the range of the kernel window width, the fixed kernel window width often leads to the loss of the target.

[0007] The CAMSHIFT (Continuous Adaptive Mean-Shift) method, as a continuous adaptive Mean Shift, can effectively solve the problem of target deformation by automatically adjusting the size of the kernel window and the size of the tracked target in the image, but its method It also converges to a local maximum, and does not make a judgment on the similarity of the target

When the search window of CAMSHIFT contains multiple candidate models with similar features, the CAMSHIFT method cannot distinguish whether there is interference, and the problem of insufficient tracking accuracy often occurs

Moreover, when the target moving object slowly passes through obstacles or performs fast and irregular movements, the CAMSHIFT method will easily fail, resulting in loss of target tracking

At the same time, when the target has a large acceleration or is blocked at an instant, the CAMSHIFT tracking method is prone to failure

Images